Five out seven extant species of sea turtles are listed as either critically endangered or endangered in the IUCN Red List of Threatened Species in 2009. Threats potentially affecting the sea turtle populations exist both on the land (e.g. degradations in nesting beaches) and in the ocean (e.g. by-catch, pollution etc.). An extensive number of sea turtle studies have been conducted around the nesting beaches where mature females land to nest, accumulating information of nesting females and hatchlings. However, there is only limited information for other life stages of sea turtles, such as males and pre-mature turtles. Therefore, understanding biology of sea turtles is essential part for proper conservation of these endangered animals.

Sanriku, the northern Pacific coast of Japan, is known as one of the most nutrient-rich fishing grounds in the Japan. Although it is located far beyond north from the northern limit of nesting grounds of sea turtles, some turtles were often sighted by local fisherman during summer. Therefore, it is suggested that Sanriku coastal area provide seasonal habitat for sea turtles.

In the present study, firstly, 1) the demographic survey was conducted at Sanriku coastal area, the northern Pacific coast of Japan, and the occurrence of sea turtles were reported. Secondly, to reveal underwater activities of loggerhead turtles in the study site, 2) three-dimensional movements were reconstructed by using multi-sensor data loggers. Then, 3) orientation behaviours during directional travel and 4) mid-water foraging behaviours were discussed.

1. Demographics of sea turtles at the northern Pacific coast of Japan

During 2005 and 2009, bycatch incidents within the range of approximately 40 km radius from the International Coastal Research Center, Ocean Research Institute, The University of Tokyo (39°21.05N, 141°54'05E) were reported. During the study periods, a total of 120 loggerhead turtles (Caretta caretta) and 39 green turtles (Chelonia mydas) were incidentally captured in the set net. Bycatch incidents were reported in the periods between June through October for loggerhead turtles, and between July through October for green turtles, suggesting the turtles seasonally migrate to the study site. The size of loggerhead turtles collected ranged from 49.5 to 88.4 cm in SCL (mean ± s.d. = 70.4 ± 6.4 cm, N = 120), which was relatively smaller than the size of adult females nesting in Japan. Therefore, it is assumed that loggerhead turtles migrating to the study site are either immature or mature turtles. On the other hand, the size of green turtles ranged from 40.7 to 85.6 cm (mean ± s.d. = 48.6 ± 10.9 cm, N = 39) in SCL, which was considerably smaller than nesting turtles, suggesting that green turtles in the study site are immature. In addition, there was a conspicuous extension in tail was observed in 4 loggerhead turtles (i.e. SCL > 70 cm, TL > 30 cm), of which were indicative of males. Therefore, it is suggested that Sanriku coastal water, the northern Pacific coast of Japan, provided seasonal habitats for immature and mature loggerhead turtles, including males, and immature green turtles.

2. Estimation of 3-D movements

Uses of animal-borne recorders are often useful to study underwater activities of marine animals. Diving behaviours of sea turtles have been studied 2-dimensionaly (i.e. time and depth). However, as turtles live in the 3-dimensional world, reconstruction of 3-D movements is required to understand their underwater activities, such as orientation and foraging behaviours. In the present study, 3-D movements were estimated for the first time for sea turtles by integrating locomotion vectors (i.e. dead-reckoning), which were computed from data obtained by a multi-sensor data logger (i.e. swim speed, depth, tri-axis geomagnetisms and tri-axis accelerations). The resulting 3-D movements successfully showed fine-scale underwater movements, which could not be detected in the traditional 2-D dive profiles. However, unrealistic fluctuations associated with strokes were observed in time-series heading data computed from tri-axis geomagnetisms and accelerations. It was suspected that such fluctuations in heading associated with stroking were artifacts derived from frequency-based filters, which were widely used in analysis of acceleration data. Although the fluctuations in heading were observed from all turtles, the effects of such fluctuations in 3-D paths were obscured at the scales of metres. In addition, as dead-reckoning could not include effect of drift by external forces, such as ocean current, the resulting 3-D movements represent how turtles moves in the water, but not the true ground position. Take account of all these precautions inherent to data analysis, a total of 213.7 hours of 3-D movements of 12 loggerhead turtles were estimated ready for the further analysis.

3. Orientation behaviours during directional travel

A total of 118 hours of 3-D movements from 8 turtles over 9 deployments were analysed to examine temporal changes during directional travel. Turtles maintained straight-line courses (straightness index _> 0.95) during 41 % of the total duration (i.e. 'travelling periods'). During travelling periods, turtle swam continuously, maintaining unidirectional heading throughout dives (mean directionality r ± s.d. = 0.99 ± 0.44, N = 2216) whereas remarkable changes in heading were occasionally observed at the surface (mean directionality r ± s.d. = 0.94 ± 0.13, N = 691). Despite highly directional movements during dives, travel direction tended to shift by the end of dives lasting tens of minutes. Such deflections in travel directions seemed to be compensated during the subsequent surfacing periods because changes in travel direction arising during dives were negatively related to that arising during subsequent surfacing periods. Hence, remarkable changes in heading at the surface could be interpreted as direction-searching behaviour. The results suggested that turtles undertaking directional travel were more dependent on directional information available at the surface although further studies are required to determine the particular source(s) of information used by the turtles.

4. Mid-water foraging behaviour

To describe foraging behaviours of free-ranging turtles, a total of 197.5 hours of 3-D movements of 12 turtles over 13 deployments were analysed. In some deployments, animal-borne video cameras ('Crittercam') were used to record the prey items. From a total of 23.8 hours of video data, a total of 71 foraging events were identified. In 67 out of 71 events, turtles fed on gelatinous prey while they were actively swimming either in the mid-water or near the sea surface. Only one turtle performed 3 benthic foraging events. By examining video and 3-D data together, it was found that most of mid-water foraging events shared a common feature having a certain period of deceleration (range of deceleration duration = 3 - 66 sec). Based on this characteristic, potential mid-water foraging events (i.e. 'presumptive events') were extracted from 3-D data. As a result, a total of 323 presumptive foraging events were extracted and the detection rate of events related to mid-water foraging was 72.3 %. By combining the true and the presumptive events, the rate of mid-water foraging events was estimated as 2.0 times h-1. Although loggerhead turtles are generally considered as benthic feeder according to digestive tract analysis, the result suggested that the previous study underestimate the importance of easily digestible gelatinous prey in the diet of loggerhead turtle. In addition, in some events, turtles changed its direction toward the prey 5.2 ± 5.2 m (mean ± s.d.) before reaching to the prey. However, such turning point tended not be exist in night-time events, suggesting that turtles used visual cue as the primary source of information in prey finding.

5. Summary and future perspective

The present study showed that some loggerhead and green turtles periodically migrate to Sanriku coastal water, the northern Pacific coast of Japan, where no nesting ground exists in proximity. The occurrences of turtles were restricted to the period between early summer to autumn, suggesting that the study site provide seasonal habitat for immature and mature loggerhead turtles, and immature green turtles. Some mid-water foraging events were confirmed when underwater activities of sea turtles at the study site were examined by using 3-D and video loggers. As being famous for its fertility, it is suggested that the Sanriku coastal water potentially provided important summer foraging grounds for turtles. In addition, by examining 3-D movements, the present study showed the turtles had ability to maintain unidirectional heading while directional travelling although turtles were more dependent on the directional cues available at the surface. Furthermore, the possibility that turtle use visual cues while feeding on gelatinous prey in the mid-water was presented.

These findings about basic biology of sea turtles would be important as a first step to understand their strategies to maximise survival in the oceanic environment. In addition, information about the distribution and underwater activities of sea turtles at the northern edge of their habitat would provide important knowledge required for proper conservation of sea turtles.

本論文は6章からなり、第1章では序論としてアカウミガメ(Caretta caretta)の生活史、分布および現在の状況について述べている。これまでアカウミガメの研究は、主に産卵場周辺において産卵上陸を行う成体メスや孵化幼体を対象に行われており、性成熟に達する前の亜成体や産卵上陸を行わないオス個体の知見は著しく不足しているのが現状であり、摂餌行動や海洋中での潜水行動になどについての研究は皆無であった。日本の北部太平洋沿岸域(三陸沿岸域)では、毎年、夏季に未成体や成体のアカウミガメが沿岸域の定置網にしばしば羅網されている。本研究では、定置網に羅網されたこれらのアカウミガメに、世界最先端のデータ ロガー(3MPD3GT)やビデオカメラ(クリターカム)を装着し、潜水行動を調査した。このデータロガーでは深度、水温、速度、重力は1秒間隔で、加速度は32HZで測定される、加速度の解析から体の3次元の姿勢や前肢のストロークの周期などを知ることができる。第2章では、三陸沿岸域において、2005年から2009年にかけて、東京大学海洋研究所国際沿岸海洋研究センター(39°21.05N, 141°54'05E)を拠点に、半径約40 km圏内に設置された定置網を対象に混獲状況の調査を実施した。その結果、合計117個体のアカウミガメの混獲を確認した。アカウミガメの混獲が確認されたのは6 - 10月であったことから、アカウミガメは季節的な回遊を行っていると考えられた。当調査地におけるアカウミガメのSCL(標準直甲長)は49.5 - 88.4 cm (平均±s.d. = 70.4 ± 6.4 cm)で、これは日本沿岸で産卵上陸を行う成体メスと比較して小さいことから、当海域に来遊するアカウミガメは主に亜成体から成体であると推測された。定置網に羅網した個体については、生物学的調査を行った後に標識を装着し、自然に戻した。同じ場所に再度戻ってくる個体もいるが、多くの個体は三陸沿岸より南の海域で再補されたことから、三陸沿岸は本種の北限に近いのではないかと類推された。第3章では、データロガーを用いてアカウミガメの潜水行動について三次元移動経路の構築する手法について述べた。この手法を用いて、アカウミガメ12個体から合計213.7時間分の三次元経路のデータを得た。このデータを解析することにより、本種の代表的な4つの潜水行動を識別し、その特徴について詳細に記述した。第4章では、三次元移動経路を用いてアカウミガメの時々刻々の遊泳方向の変化を調べ、移動を行うアカウミガメの定位について調査した。調査時間の41%の期間でアカウミガメは直線的に移動を行っていた。移動中、本種は水中では方向を変化させずに一定の方向を游泳するのであるが、水面に浮上する際にはしばしば頭を360度回転させていることが明らかになった。データロガーで得られたこの特徴的な行動は、アカウミガメに一緒に装着したビデオ記録からも検証された。この行動は昼夜で確認されていることから、その理由について議論された。第5章では、三次元移動経路に加え、動物装着型ビデオカメラで撮影したビデオの記録を用いて解析を行い、アカウミガメの浮遊性動物を捕食する摂餌行動について調査した。1時間あたりの摂餌イベント回数は2.0回と推定された。排泄物や消化管内容物を調べた既往研究によると、アカウミガメは一般に底性動物を摂餌すると考えられていたが、本研究の結果より、クラゲ等の消化速度の早い浮遊性動物の摂餌量が低く見積もられている可能性が示唆された。約7割の摂餌イベント時、アカウミガメは餌生物の5.3±5.2 m(平均±s.d., N = 263)手前で方向転換した後、餌生物に向かって直進して餌生物を捕獲した。しかし夜間のイベントでは、このような方向転換がみられないイベントが多い傾向がみられた。これらのことから、アカウミガメは視覚に頼って浮遊性の餌生物を発見している可能性が示唆された。最後に第6章では、得られた結果をまとめて総合的に討論し、アカウミガメの生息域の北限に近い三陸沿岸域における本種の保全に対する今後の展望が述べられている。

なお、本論文第4章の内容は、Sea turtles compensate deflection of heading at the sea surface during directional travelというタイトルでThe Journal Experimental Biology第212巻に投稿済みである。本章は、佐藤克文氏、Kyler J. Abernathy氏、Greg J. Marshall氏との共同研究であるが、論文提出者が主体となってデータ収集および解析を行ったもので、論文提出者の寄与が充分であると判断する。

以上、本研究は、海洋動物の潜水行動と環境選択の解明に極めて有意義な知見を得たことから、学術上、応用上貢献することが少なくない。よって審査委員一同は本論文が博士(環境学)の学位論文として価値があるものと認めた。